04 — Visualizations¶

All analysis charts:

  1. Scatter + Regression: Performance vs. Salary per position
  2. Bar Charts: Top bargains and overpaid per position
  3. Box Plots: Salary distribution by position
  4. Year-over-Year Trends: Value efficiency over time
  5. Team Heatmap: Best-value franchises by position
  6. Rookie vs. Veteran: Contract value distributions
  7. Player Trajectories: Individual deep dives
In [1]:
import sys
sys.path.insert(0, '..')

import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore', category=FutureWarning)

from src.data_loader import DATA_DIR
from src.viz import (
    scatter_performance_vs_salary,
    bar_chart_top_values,
    boxplot_salary_by_position,
    year_over_year_trends,
    team_value_heatmap,
    rookie_vs_veteran_comparison,
    player_trajectory,
    all_positions_scatter_grid,
)

import plotly.io as pio
pio.renderers.default = 'notebook'

pd.set_option('display.max_columns', 50)

E:\Python\Python38-32\lib\site-packages\numpy\_distributor_init.py:30: UserWarning: loaded more than 1 DLL from .libs:
E:\Python\Python38-32\lib\site-packages\numpy\.libs\libopenblas.D6ALFJ4QQDWP6YNOQJNPYL27LRE6SILT.gfortran-win32.dll
E:\Python\Python38-32\lib\site-packages\numpy\.libs\libopenblas_v0.3.21-gcc_8_3_0.dll
  warnings.warn("loaded more than 1 DLL from .libs:"
In [2]:
scored = pd.read_parquet(DATA_DIR / 'scored.parquet', engine='fastparquet')
print(f"Loaded: {scored.shape[0]:,} player-seasons")
print(f"Positions: {sorted(scored['pos_group'].unique())}")
print(f"Seasons: {sorted(scored['season'].unique())}")

Loaded: 9,041 player-seasons
Positions: ['DB', 'DL', 'K', 'LB', 'OL', 'QB', 'RB', 'TE', 'WR']
Seasons: [2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023, 2024]

5a. Scatter: Performance vs. Salary¶

Interactive scatter plots per position group. Hover for details, outliers labeled.

In [3]:
# QB scatter (interactive)
fig = scatter_performance_vs_salary(scored, 'QB', highlight_threshold=1.5)
if fig: fig.show()
In [4]:
# WR scatter
fig = scatter_performance_vs_salary(scored, 'WR', highlight_threshold=1.5)
if fig: fig.show()
In [5]:
# RB scatter
fig = scatter_performance_vs_salary(scored, 'RB', highlight_threshold=1.5)
if fig: fig.show()
In [6]:
# DL scatter
fig = scatter_performance_vs_salary(scored, 'DL', highlight_threshold=1.5)
if fig: fig.show()
In [7]:
# All positions grid (static)
fig = all_positions_scatter_grid(scored, highlight_threshold=1.5)
fig
Out[7]:

5b. Top/Bottom Value Rankings¶

In [8]:
# Top 15 Bargains — All positions
fig = bar_chart_top_values(scored, n=15, chart_type='bargains')
fig
Out[8]:
In [9]:
# Top 15 Overpaid — All positions
fig = bar_chart_top_values(scored, n=15, chart_type='overpaid')
fig
Out[9]:
In [10]:
# QB Bargains
fig = bar_chart_top_values(scored, pos_group='QB', n=15, chart_type='bargains')
fig
Out[10]:
In [11]:
# QB Overpaid
fig = bar_chart_top_values(scored, pos_group='QB', n=15, chart_type='overpaid')
fig
Out[11]:

5c. Salary Distribution by Position¶

In [12]:
fig = boxplot_salary_by_position(scored)
fig
Out[12]:

5d. Year-over-Year Trends¶

In [13]:
fig = year_over_year_trends(scored)
fig.show()

5e. Team Value Heatmap¶

In [14]:
# Last 3 seasons
fig = team_value_heatmap(scored, seasons=[2022, 2023, 2024])
fig
Out[14]:
In [15]:
# Full history
fig = team_value_heatmap(scored)
fig
Out[15]:

5f. Rookie vs. Veteran Value¶

In [16]:
fig = rookie_vs_veteran_comparison(scored)
fig
Out[16]:

5g. Player Deep Dives¶

Interactive trajectory plots for specific players.

In [17]:
# Patrick Mahomes
fig = player_trajectory(scored, 'Mahomes')
if fig: fig.show()
In [18]:
# Josh Allen
fig = player_trajectory(scored, 'Josh Allen')
if fig: fig.show()
In [19]:
# Derrick Henry
fig = player_trajectory(scored, 'Derrick Henry')
if fig: fig.show()
In [20]:
# Aaron Donald
fig = player_trajectory(scored, 'Aaron Donald')
if fig: fig.show()
In [21]:
# Justin Jefferson
fig = player_trajectory(scored, 'Jefferson')
if fig: fig.show()

Custom Player Lookup¶

Change the player name below to explore any player.

In [22]:
PLAYER_NAME = 'Lamar Jackson'  # <-- Change this

fig = player_trajectory(scored, PLAYER_NAME)
if fig: fig.show()

# Show detailed stats
player_data = scored[scored['player_name'].str.contains(PLAYER_NAME, case=False, na=False)]
if len(player_data) > 0:
    cols = ['season', 'player_name', 'pos_group', 'recent_team',
            'apy_cap_pct', 'performance_zscore', 'salary_zscore',
            'value_score', 'value_percentile']
    cols = [c for c in cols if c in player_data.columns]
    display(player_data[cols].sort_values('season'))
season player_name pos_group recent_team apy_cap_pct performance_zscore salary_zscore value_score value_percentile
354 2018 Lamar Jackson QB BAL 0.013 -0.118734 -0.789203 0.670469 77.777778
355 2019 Lamar Jackson QB BAL 0.013 1.482478 -0.789203 2.271681 100.000000
356 2020 Lamar Jackson QB BAL 0.013 0.900254 -0.789203 1.689457 97.575758
1163 2020 Lamar Jackson DB NYJ 0.001 -0.428535 -0.666978 0.238443 64.313082
357 2021 Lamar Jackson QB BAL 0.013 0.280714 -0.789203 1.069917 90.707071
1857 2022 Lamar Jackson DB 2TM 0.004 -1.003134 -0.503259 -0.499876 19.704433
358 2023 Lamar Jackson QB BAL 0.144 1.164035 0.999077 0.164958 57.171717
359 2024 Lamar Jackson QB BAL 0.231 1.928940 2.186713 -0.257772 37.171717